/ /

in ih

If

)r

1,

0

IS

‘Y

)-

k

- It I. -

1

1

I

ORIGINAL ARTICLES 31Phosphorus Magnetic Resonance Spectroscopy of the Frontal and Parietal Lobes in Chronic Schizophrenia

Raymond F. Deicken, Giovanna Calabrese, Edward L. Merrin, Dieter J. Meyerhoff, William P. Dillon, Michael W. Weiner, and George Fein

In vivo ”Phosphorus magnetic resonance spectroscopic imaging (31P MRSI) was pegormed on 20 chronic schizophrenic patients and 16 normal controls to determine i f there were specific changes in high energy phosphorous and phospholipid metabolism in the frontal lobes of schizophrenic patients. Phosphorous metabolites were assessed in each of the le& and right frontal as well as the left and right parietal lobes. Frontal lobe phosphorous metabolites were also correlated with severity ofpsychiatric symptomatology as assessed by the Brief Psychiatric Rating Scale (BPRS). Schizophrenics demonstrated higher phosphodiesters (PDE) and lower phosphocreatine (PCr) in both the lef and rightfrontal regions compared to controls. There was also lower left frontal inorganic phosphate (Pi) in the schizophrenic group. No group differences were noted in the left or right parietal regions. In addition, right frontal PDE and right frontal PCr were highly correlated with the hostility-suspiciousness and anxieq-depression subscales of the BPRS. This study provides further support for altered frontal lobe phosphorous metabo- lism in schizophrenia.

Key Words: Magnetic resonance spectroscopy, schizophrenia, phosphorous, phospholipid, frontal lobe. brain metabolism

Introduction Studies of schizophrenia have implicated the frontal lobe as a site of impaired functioning. Investigations of regional metabolic activity in the brain of patients with schizophre- nia using positron emission tomography (PET) have re-

From the Magnetic Resonance Unit (RFD, DIM. MWW) and the Psychiatry Service, Department of Veterans Affairs (DVA) Medical Center, San Francisco, CA (RFD, ELM. GF): the Departments of Radiology (DJM, WPD. MWW). Medl- cine (MWW), and Psychiatxy (RFD. ELM, GF), University of California San Francisco, LA; and the Department of Psychiatry, H.S. Raffaele InstiNte, Uni- vmity of Milan, Italy (GC).

Address reprint requests to Raymond F. Deicken. MD, Psychiatry Service, Veterans AffairsMedicalCenter(l16-N),415OClement St.SanFrancisco.CA94121.

This research was presented at the annual meeting of the Society of Biological Psychiatry. San Francisco. CA. May 19-23.1993.

Received July 7,1993; revised November 20,1993.

0 1994 Society of Biological Psychiatry

ported a relative decrease of glucose metabolism and cere- bral blood flow in frontal brain regions compared to posterior brain regions, a phenomenon termed “hypofron- tality” (Buschbaum et a1 1982; Farkas et al 1984; Wolkin et a1 1985). Other PET studies in which tracer uptake took place during performance of tasks designed to potentially engage frontal cortex have also demonstrated reduced acti- vation of the frontal lobe in schizophrenia (Volkow et al 1987; Weinberger et al1986).

More recently, Pettegrew and colleagues (1991) utilized nonlocalized in-vivo 31Phosphorous (”P) magnetic reso- nance spectroscopy to study the left dorsal prefrontal cortex in drug-naive, first-episode schizophrenic patients. The schizophrenic group demonstrated significantly increased

0006-3223/94/507.00

504 BlOL PSYCHIATRY 1994;36:503-510

phosphodiesters (PDE) and decreased phosphomonoesters (PME) compared to controls. The authors suggested that these findings were consistent with decreased synthesis and increased breakdown of membrane phospholipids, although other interpretations are possible. The schizophrenic group was also noted to have increased adenosine triphosphate (ATP) and decreased inorganic orthophosphate (Pi), which was interpreted to be consistent with hypoactivity of the dorsal prefrontal cortex. Another single volume 31P mag- netic resonance spectroscopy study of the left prefrontal cortex in medicated schizophrenics demonstrated de- creased PME and increased phosphocreatine (PCr) and Pi (Williamsonetall991). Keshavanetal(1989)alsoreported that after 4 weeks of neuroleptic treatment, first-episode, neuroleptic-naive schizophrenics who initially demon- strated decreased left prefrontal cortex PME and alpha-ATP as well as increased PDE and F'Cr compared to controls continued to show increased PDE, further increases in PCr, as well as further decreases in PME and alpha-ATP. They interpreted their findings as suggesting accelerated break- down of membrane phospholipids and decreased utilization of PCr in the dorsal prefrontal cortex of schizophrenics.

The present study was conducted to replicate and extend these findings by determining if the alterations in left frontal lobe phosphorous metabolites as detected by magnetic reso- nance spectroscopy are specific to the frontal lobe. Mag- netic resonance spectroscopic imaging (MRSI) was utilized to measure phosphorous metabolites in both the frontal and parietal regions of schizophrenic patients and normal con- trols. The major advantage of 31P MRSI over the single volume MRS techniques used in the previously described studies is that it allows the study of phosphorous metabolites from several different brain regions at the same time. A second objective of this study was to'determine if there was an association between frontal lobe phosphorous metabo- lites and the severity of psychiatric symptomatology.

Materials and Methods

Subjects Twenty men who were schizophrenic patients (mean -C SD age = 39.2 2 5.8 years) and 16 men who were control subjects (mean ? SD age = 40.3 ? 11 .O years) gave in- formed consent for participation in the study. The diagnosis of chronic schizophrenia was confirmed using the Struc- tured Clinical Interview for DSM-111-R patient edition (SCID-P) (Spitzer and Williams 1985) by two psychiatrists (ELM and RFD) who arrived at a consensus diagnosis in each case. Control subjects were also assessed using the SCID nonpatient edition (SCID-NP). All schizophrenic pa- tients and control subjects were right handed. The demo- graphic and clinical characteristics of the schizophrenic

R.F. Deicken et d

patients are shown in Table 1. All the time of the study, six patients were neuroleptic free, seven patients were taking neuroleptics only; three patients were taking a neuroleptic plus diphenhydramine; and four patients were taking a neuroleptic plus benztropine mesylate. None of the schizo- phrenic patients had a history of head injury, organic mental disorder, neurological disorder, major mood disorder, or anxiety disorder. One of the patients had clinically signifi- cant alcohol abuse 10 months prior to the study and one patient had clinically significant cannabis abuse 11 months prior to the study. The remaining patients had no significant alcohol or substance abuse for at least 1 year prior to the study. The control subjects consisted of nine Caucasian, 3 Black, 2 Asian, and 2 Hispanic men with a mean 5 SD years of education of 13.6 ? 1.5. None had any history of signifi- cant medical illness, head injury, neurological disorder, psychiatric disorder, or clinically significant alcohol or sub- stance abuse. There were no significant group differences between patients and controls for age or education. The clinical symptomatology of each patient was assessed by two psychiatrists (ELM and RFD) using the 18-item Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham 1962). Interrater carelation coefficient was 0.9 1.

MRSI Methods All MRSI studies were performed at the San Francisco Department of Veterans Affairs Medical Center's Magnetic Resonance Unit on a Philips Gyroscan S 15 MRI/MRS sys- tem operating at 2 Tesla. A standard imaging saddle-type proton head coil was used for MRI. TI-weighted sagittal (7 slices, 7.1 mm thick, 1.2 mm gap, TR = 600 msec, TE = 30 msec) and T,-weighted axial (16 slices, 7.1 mm thick, 1.2 mm gap, TR = 2000 msec, TE = 30 and 80 msec) multi-slice images were obtained on each subject. These MRI images were subsequently evaluated by a board-certified neuro- radiologist (WPD) for the presence of abnormalities or atrophy. The axial slices were angulated parallel to the canthomeatal plane observed on the sagittal slices to give a consistent anatomical perspective, facilitating comparisons between patients and controls. The Bo magnetic field was then shimmed (nonlocalized) using the proton imaging coil and Philips computer-controlled optimization for a water resonance linewidth of less than 15 Hz. The subject was then removed from the magnet bore while the 31P MRSI study was prepared. For accurate repositioning of the sub- ject's head, a vacuum-assisted head holder, padded straps, and surface anatomical markings (forehead and temple) were employed. An inductively coupled, high-pass, quadra- ture birdcage head coil was used to provide homogeneous RF excitation and detection. The MRSI procedures and experimental parameters have been previously described (Maudsley et a1 1990; Hugg et a1 1992). A spin-echo se- quence (TR = 350 msec, TE = 3.5 msec) was utilized. The

t LI

3lp MRS in sc

Table 1. Den

Patient

1 2 3 4 5 6 7 8 9

10 I 1 12 13 14 15 16 17 18 19 20 Mean Z SD

MRSI ra\k with 1728 256 samp sitions we

MRSI transform VAX wo each of 1 voxel sizi line broa The MR! display. 1

ated don The high was usec each sub. addition; Figure 1

3'P sp processi cuse, Nc signal fi

convolu Spectral PCr, wh age of tc @ATP to best I unlike t bution f

31P MRS in Schizophienia BlOL PSYCHIATRY 505 lW:36:5O3-5 IO

Table 1. Demographic and Clinical Characteristics of Schizophrenic Patients Education Diagnosis Yeanof BPRSscore Neuroleptic

medication Patient Age Race (Yean) (DSM-111-R) illness (tow

1 35 Asian 12 Disorganized 16 52 Chlorpromazine 500 mglday 2 38 Caucasian 12 Paranoid

50 None x 7 days 4 42 Black 13 Undifferentiated 19 34 Haloperidol 20 mglday 5 40 Caucasian 7 Undifferentiated 16 41 None x 6 days 6 46 Black 7 53 Caucasian 8 49 Black 14 Disorganized 22 44 None 9

10 11 12 39 Caucasian 12 Disorganized 19 55 Clozapine 500 mg/day

Caucasian 13 Disorganized 25 61 Clozapine 150 mg/day 13 40 14

13 Paranoid 2 28 Haloperidol 10 mg/day 15 31 Caucasian 16 40 Black 12 Paranoid 15 43 None x 1 days

56 None x 7 days 17 40 Caucasian 13 Undifferentiated 20 Undifferentiated 12 44 Thiothixene I O mg 18 32 Caucasian 15

19 38 Caucasian 13 Undifferentiated ' 10 40 Thiothixene 10 mg 20 39 Black 15 Paranoid 5 27 None Mean i SD

4 41 Perphenazine I6 mg/day 3 34 Black 13 Paranoid 20

12 Undifferentiated 1.5 51 Remoxipride 90 mg/day 12 Undifferentiated 27 35 Haloperidol 20 mg/day

36 Asian 14 Undifferentiated 11 56 Chlorpromazine 550 mg/day 43 Caucasian 12 Undifferentiated 10 54 Perphenazine 32 mg/day 40 Caucasian 13 Paranoid 2 39 Chlorpromazine 175 mg/day

30 Black 13 Undifferentiated I 43 Perphenazine 24 mg/day

12.6 i 1.6 13.6 t 7.8 45.8 f 9.5 39.2 i 5.8

MRSI raw data volume consisted of a field of view of 27 cm3 with 1728 phase-encoded half-echo time signals, each with 256 samples and a bandwidth of 3 kHz. Sixteen data acqui- sitions were averaged to give an acquisition time of 86 min.

MRSI data volumes were reconstructed using Fourier transforms with apodization and zero-filling on a micro- VAX workstation to yield 32 x 32 interpolated voxels in each of 16 transverse planes (1.7 cm thick). The effective voxel size was about 25 cm3. A I O Hz exponential spectral line broadening was applied to optimize signal-to-noise. The MRS images were interpolated to 64 x 64 voxels for display. A reference image of the total 31P signal was gener- ated along with individual phosphorous metabolite images. The higher resolution of the spatially registered MR images was used to select two voxels in comparable locations for each subject in the right and left frontal lobes as well as two additional voxels in the right and left parietal lobes (see Figure 1).

31P spectra from these voxels were fit by NMRl data processing software (New Methods Research, Inc., Syra- cuse, New York) on a SUN 3/60 workstation. The broad signal from the less-mobile metabolites was removed by convolution difference and baseline flattering procedures. Spectral peaks were fit to Gaussian lineshapes except for PCr, which was fit to a Lorentzian lineshape. The percent- age of total phosphorous signal for PME, Pi, PDE, PCr, and P-ATP was calculated. The @-ATP resonance was selected to best represent the ATP concentration in tissue, because unlike the y and a resonances, it is free from signal contri- bution from other phosphate-containing metabolites such as

adenosine diphosphate and nicotinamide adenine dinucleo- tide phosphates (Glonek et al 1982). Intracellular pH was derived from the chemical shift of Pi referenced to the chemical shift of PCr in ppm (Petroff et a1 1985). Intraclass correlation coefficients for both interoperator and intra- operator reliability of NMR 1 spectral fitting have been determined in our laboratory for 31P MRSI spectra from regional white matter voxels in 16 normal control subjects (C. Husted, unpublished data). The interoperator correla- tion coefficients were as follows: 0.45 for PME, 0.93 for Pi, 0.79 for PDE, 0.88 for PCr, 0.73 for P-ATP. The intraopera- tor correlation coefficients were as follows: 0.75 for PME, 0.99 for Pi, 0.86 for PDE, 0.95 for PCr, and 0.95 for P-ATP. This data indicates that the reliability for a single operator is greater than for two operators and that it is preferable to have one operator evaluate all spectra in a given set of control and disease 31P data. Therefore, to minimize the inter-operator variability of spectral processing in the present study, a single operator (RFD) processed all spectra which were blindly coded.

Statistical Analysis The a priori planned comparisons involved examination of the group by region interactions for PDE and PME with PDE hypothesized to be higher and PME lower in the frontal lobes of schizophrenic patients. Multivariate repeated mea- sures analysis of variance (MANOVA) was used for the major analyses. The dependent variable for each metabolite was its percentage of the total phosphorous signal, group was the between-subjects factor, and side (left versus right)

506 BIOL PSYCHIATRY i!W;36:503-5 10

and region (frontal versus parietal) were the within-subjects repeated measures factors. The a priori planned compari- sons involving PDE and PME were Bonferroni corrected for two comparisons with the significance set t op < 0.025. The remaining phosphorous metabolites were analyzed in an exploratory fashion without correction for multiple com- parisons. Regression analysis was used to examine the asso- ciation between frontal lobe phosphorus metabolites and severity of psychiatric symptomatology as assessed by the total BPRS scores and five subscale scores of the BPRS as follows: ( 1) withdrawalhetardation (BPRS items for emo- tional withdrawal, motor retardation, blunted affect, disor- ientation), (2) hostility/suspiciousness (BPRS items for hostility, suspiciousness, uncooperativeness). (3) thinking disturbance (BPRS items for conceptual disorganization, grandiosity, hallucinatory behavior, unusual thought con- tent), (4) anxiety/depression (BPRS items for somatic concern, anxiety, guilt feelings, depressive mood), and (5) agitation/excitement (BPRS items for tension, excite- ment). Bonferroni correction for six correlations examined resulted in a significance criterion o f p < 0.008.

Results No structural abnormalities or atrophy were noted on any of the MR images. Table 2 shows the percentage of phospho- rous metabolites and pH for the frontal and parietal lobes as well as the MANOVA results. There was a significant re- gion effect with higher PDE (F = 8.53, df = 1,34, p = 0.006) and lower ATP ( F = 8.65, df = 1,34, p = 0.006) in parietal compared to frontal regions. In addition, a significant region bygroupinteractionforPDE(F= 12.9,df= 1,34,p=O.001) indicated higher frontal PDE in the schizophrenics com- pared to control subjects. There were no significant differ- ences in PME between groups in either the frontal or parietal regions. For PCr, there was a region by group interaction (F = 4.51, df = 1,34, p = 0.04) with schizophrenics having

Figure 1. MR images from a control subject showing the location of se- lected voxels in the left and right frontal as well as the left and right parietal lobes.

lower PCr in the frontal lobes than the controls. A region by side by group interaction was also present for Pi (F = 5.60, df = 1,34, p = 0.024) with left frontal regions having rela- tively lower Pi than right frontal regions in the schizo- phrenic group.

Regression analysis of the frontal and parietal lobe phos- phorous metabolites and total BPRS scores revealed a posi- tive correlation with right frontal PDE ( r = 0.47, p = 0.035) and a negative correlation with right frontal PCr ( r = -0.42, p = 0.07). In addition, the hostility-suspiciousness subscale of the BPRS was positively correlated with the right frontal PDE (r = 0.72, p = O.OOO4) and negatively correlated with the right frontal PCr (r = -0.66. p = 0.001). The anxiety-de- pression subscale was also positively correlated with right frontal PDE (r = 0.59, p = 0.006) and negatively correlated with right frontal (r = -0.47, p = ,038). The hostility- suspiciousness subscale and anxiety-depression subscales were also positively correlated with each other (r= 0 . 6 6 , ~ = .002), however. There were no significant correlations of the frontal lobe phosphorous metabolites and the with- drawal-retardation, thinking disturbance, or agitation-ex- citement subscales of the BPRS. There were also no signifi- cant correlations between the frontal and parietal lobe phosphorous metabolites and either years of illness or neur- oleptic dosage in chlorpromazine equivalents.

Discussion The first major finding of this study is that compared to control subjects, PDE was significantly higher in both the right and left frontal lobes of medicated schizophrenic pa- tients. In contrast, there was no significant difference in PDE between the schizophrenics and controls in the parietal lobes. In the normal controls, PDE was lower in the frontal compared to the parietal lobes, but in the schizophrenic group, there was no significant difference between frontal and parietal PDE. The PDE findings in the schizophrenic

3 l P MRS in Schizophienia BlOL PSYCHIATRY 507 19!34;36:503-5 I O

by 50, la- IO-

os- )si- 3 5 ) 42. :ale ntal vith -de- ight ated ity- ales . p= IS of vlth- -ex- :nifi- lobe ieur-

ed to :h the ic pa- s e in trietal rontal uenic iontal hrenic

Table 2. Metabolite Percentage of Total Phosphorous Signal and MANOVA Analysis MANOVA Mean Values -t SD

Controls Schizophrenics (n = 16) (n = 20) Source F P

PME

Pi

PDE

PCr

P-ATP

PH

RF LF RP LP

RF LF RP LP

RF LF RP LP

RF LF RP LP

RF LF RP LP

RF LF RP LP

13.0 2 3.9 1 3 3 2 3.2 14.4 2 2.8 14.9 i 5.0

6.6 Z 2.5 6.4 2 3.3 5.4 i 2.3 5.1 i 2.2

27.1 I 3.7 28.0 z 5.3 30.8 f 3.5 32.1 z 4.9

21.4 z 3.2 20.6 z 3.4 20.9 I 2.2 18.9 f 3.3

9.4 t 7.6

8.1 i 1.7 9.8 z 2.7

9.0 t 2.6

6.97 I 0.28 7 . 0 3 i 0.12 7.03 2 0.09 7.04 z 0.09

13.5 t 4.0 13.5 2 3.9 14.8 2 3.1 13.6 z 4.4

6.7 I 3.4 5.2 i 2.1 4.6 2 1.7 6.7 z 2.8

30.3 z 3.0 31.3 z 3.4 31.1 i 3.2 29.7 2 4.7

19.0: 3.8 20.0 i 3.9 20.1 f3.1 20.9 i 4.7

8.8 f 1.7 8.9 I 2.4 8.0 z 1.8 8 3 z 1.9

7.05 i 0.10 7.04 i 0.08 7.05 t 0.07 7.01 i 0.07

Group Region R x G Side S x G R x S R x S x G

Group Region R x G Side S x G R x S R x S x G

Group Region R x G Side S x G R x S R x S x G

Group Region R x G Side S x G R x S R x S x G

Group Region R x G Side S x G R x S R x S x G

Group Region R x G Side S x G R x S R x S x G

0.06 3.05 0.22 0.002 0.90 0.48 0.14

0.05 2.91 1.22 0.002 0.49 5.23 5.60

1.95 8.53

0.45 0.92 0.83 1.51

0.30 0.004 4.5 1 0.18 4.22 0.46 0.23

I .68 8.65 0.87 2.42 0.30 0.35 0.02

1.17 0.22 I .06 0.07 2.79 1.33 0.16

12.9

NS NS NS N S NS NS NS

NS NS NS NS NS 0.03 0.02

NS 0.006 0.00 1

NS NS NS NS NS NS 0.04 NS NS NS NS

NS 0.006 NS NS N S NS N S

NS NS NS NS NS NS NS

RF = Right Frontal Lobe, LF = Left Frontal Lobe. RP = Right Panetal Lobe. LP = Left Panetal Lobe R x G = Region by Group. S x G = Side by Group. R x S = Region by Side, R x S x G = Region b) Side hy Group. NS = nonsignificant

p u p are consistent with those of Pettegrew and associates 199 1 1 who studied the left dorslateral prefrontal cortex in

1 1 first-episode, never-medicated patients using a nonloca- hied technique. A recent report using a localized surface coil technique (Stanley et a1 1992) noted a higher left pre- frontal cortex PDE in a small sample (n = 6) of newly diagnosed. medicated schizophrenics compared to young controls but not in chronic, medicated schizophrenics com-

pared to middle-aged controls, however. The patients in our study were for the most part chronic, medicated schizo- phrenics with an average length of illness of 13.9 years. The PDE findings of these studies together raise the question of whether alterations in frontal lobe PDE metabolism persist in chronic patients treated with neuroleptic medication.

Any interpretation of changes in PDE must take into account that the PDE peak consists of contributions from

508 BlOL PSYCHIATRY 1994:36:50>510

glycerophosphorylcholine, glycerphosphorylethanolamine, and mobile phospholipids. Changes in the relative intensity of the PDE peak might be due to a true concentration differ- ence in any one of these contributing compounds, or alter- natively, to changes in either the T1 or T2 of these com- pounds. One possible interpretation suggested by Pettegrew et al(1991) is that increased frontal cortex PDE may reflect decreased PDE phosphodiesterase activity and increased phospholipase AI and A2 activity. Studies of peripheral blood seem to implicate disturbances in phospholipid me- tabolism in schizophrenia as well, although it is not clear how to reconcile such findings with specific frontal lobe phospholipid changes. For example, Gattaz and colleagues (1987, 1990) have noted significantly higher plasma, serum, and platelet levels of phospholipase A2 in schizo- phrenics compared to controls. Recently. Pangerl et a1 ( 199 1) reported increased platelet lysophosphatidylcholine (LPC) in schizophrenics compared to controls, which is consistent with the finding of increased phospholipase A2 as LPC is a toxic metabolite of phospholipase A2 action on phosphatidylcholine. LPC has been known to affect the integrity of cell membranes and second messenger systems (van den Bosch 1980, Carter et a1 1990, Oishi et a1 1990). In addition,, Horrobin and colleagues (1991) have reported significant differences from norm in the frontal cortex fatty acid composition of phosphatidylethanolamine in schizo- phrenics. These findings taken together suggest that there is increasing evidence to suspect significant disturbances in phospholipid metabolism in schizophrenia, and that in the brain, some of these disturbances may be specific to the frontal lobe.

Unlike previous studies (Pettegrew et a1 1991; William- son et a1 19911, we did not find decreased PME in the schizophrenic group. This may be due to differences in clinical populations or in-vivo spectroscopy techniques. For example, the frontal lobe VOIs in this study contained sub- stantial contributions from both gray and white matter as opposed to the previous studies that looked primarily at gray matter (left dorsal prefrontal cortex). It should also be noted that our spectral processing of PME appears to have lower interoperator and intraoperator reliability than the other phosphorous metabolites in the spectrum. The lower reli- ability of PME may be due to the fact that it resonates at the low field end of the spectrum and may be more suscep- tible to changes in phasing and baseline during spectral processing.

The second finding is that of decreased right and left frontal PCr and decreased left frontal Pi in schizophrenics compared to controls. It is important to note, however, that these findings arose from exploratory analyses that were uncorrected for multiple comparisons and must be repli- cated in future studies. Fujimoto and colleagues (1992) have also recently reported decreased PCr in the left frontal re-

R.F. Deicken et al

gion in chronic, medicated schizophrenics using 31P MRSI. The decreased PCr could be due to decreased PCr produc- tion or, alternatively, to increased PCr utilization that might occur in a state of increased frontal metabolism. In the latter scenario, tissue concentrations of PCr would be reduced in response to a shift of the creatine kinase equilibrium pro- duced by increased ADP. Our finding of decreased left frontal Pi is similar to Pettegrew et a1 (1991) who reported increased ATP and decreased Pi in never medicated patients which they interpreted as consistent with frontal hypometa- bolism (decreased ATP utilization). Some PET studies have found no hypofrontality (Sheppard et a1 1983; Kling et a1 1986; Volkow et al1986) andeven hyperfrontality (Szecht- man et a1 1988) in schizophrenia, however. Furthermore, it is not known what effect neuroleptics have on brain high energy phosphorous metabolism. PET studies have demon- strated that neuroleptics and exacerbation of psychosis fol- lowing application of dopaminergic agonists have both been shown to increase brain metabolic activity (Wolkin et a1 1985; DeLisi et a1 1985; Geraud et a1 1987). Given our mixed sample of patients both on and off neuroleptics, it is difficult to draw any definitive conclusions about the meta- bolic significance of our PCr and Pi findings.

The third finding of this study are the correlations be- tween the right frontal PDE and PCr measures and the BPRS scores. Higher right frontal PDE and lower right frontal PCr in the schizophrenic group were associated with increasing severity of overall psychiatric symptoms as measured by the BPRS, with the effect being most pronounced for the hostility/suspiciousness and anxiety/depression symptom clusters. The negative right frontal PCr correlation with the hostility/suspiciousness subscale is similar to our previous 31P MRS study of the temporal lobes in schizophrenia (Ca- labrese et a1 1992) where we found significant negative correlations between the total BPRS scores and both right and left temporal lobe PCr/ATP ratios as well as PCr. This suggests that for the right frontal and both temporal lobes, altered brain high energy phosphorous metabolism as mea- sured by PCr may reflect the degree of psychiatric symptom control. One must exercise caution when interpreting the observed correlations as our sample of schizophrenics con- sisted of patients who were either neuroleptic free or taking a variety of neuroleptics. It is possible that the symptoms assessed by the BPRS may in part represent indices of the patients' failure to respond to treatment.

In-vivo 3'P spectroscopy of psychiatric disorders is still in its infancy. In comparison with schizophrenia, elevated PDE and reduced PME in the dorsolateral prefrontal cortex have also been reported in autism (Minshew et a1 1989). Recent studies of bipolar disorder have shown elevated frontal lobe PME in both lithium treated and untreated manic patients compared to the euthymic state and normal controls (Kato et al 1991, 1992). From these studies as well

3

i t E c

S

r C

\

r

r a t F

F n

S

t*

S

S

n F

a d g ti tl h

C

S

1i

n SI

ii fl

ii

n n CI

E B

C

C

3 I P MRS in Schizophienia BlOL PSYCHIATRY 509 1994;36:503-5 10

ve a1 It- it

,

:h n- 11- th et ur is a-

e- .S 3 lg 'Y le m ie

I

1s

3-

le

is S , *- n ie

g

le

ht

1-

1s

n d X

1. d

il

d i i

11 1

as those in schizophrenia to date, it is tempting to speculate that altered frontal lobe phospholipid metabolism may play a role in the pathophysiology of several major psychiatric disorders .

With regard to the limitations of the present study, our sample size was small and our findings clearly need to be replicated in a larger population. Second, our study included different clinical subtypes of schizophrenia and patients who were neuroleptic free as well as those on a variety of neuroleptics. It is hoped that future studies with larger sample sizes will be able to address the effects of chronic neuroleptic medication on phosphorous metabolites and any differences between the clinical subtypes. Third, poten- tial frontal cortical atrophy may have had an effect on our phosphorous measurements. Although there were no quali- tative differences in frontal atrophy between the schizo- phrenics and controls, more precise quantitative measure- ments are necessary with this relatively new technology. If significant atrophy were present in the frontal lobes of the schizophrenic group, however, there would probably be noticeable reductions in all phosphorous metabolites. Fourth, the voxels selected for each subject were 25 cc and contained varying percentages of gray matter, white matter, and cerebral spinal fluid (CSF). This limits our ability to draw any conclusions about how specific our findings are to gray matter or white matter. In the future, MRI segmenta- tion software can be interfaced with MRSI to determine the percentage of gray matter, white matter, and CSF in MRSI voxels.

It is also relevant to mention the possible interrelation- ships of the MRSI phosphorous measures. The first is re- lated to the potential day-to-day instrumental variation that may be assessed by using each subject as its own control in serial studies. The second is to what degree MRSI measures in one brain region are contaminated by MRSI measures from another brain region. For the brain regions examined in this study that involved nonadjacent voxels, MRSI measures in one voxel are largely independent of MRSI measures in the other voxels. There can be some degree of covariance of MRSI measures when comparing adjacent

References Buchsbaum MS, Ingvar DH, Kessler R, et al (1982): Cerebral

glucography with positron tomography: use in normal subjects and patients with schizophrenia. Arch Gen Psychiatry 39:25 1- 259.

Calabrese G, Deicken RF, Memn EL, Schoenfeld F, Fein G, Weiner MW (1992): 31Phosphorous magnetic resonance spec- troscopy of the temporal lobes in Schizophrenia. Biof Psychia- try 32:26-32.

Carter HR, Wallace MA, Fain JN (1990): Activation of phospholi- pase C in rabbit brain by carbachol in the presence of GTP gamma S: effects of biological detergents. Biochem Biophys Acta 1054: 129-1 35.

voxels where there is some overlap of signal between voxels, however. The third involves spectral processing where the curve fitting of one metabolite affects the curve fitting of other metabolites. Because in-vivo 31P MRS spec- tra have overlapping peaks, particularly in the region in- volving PME, Pi, PDE, and PCr, there is the possibility that the curve fitting of one spectral peak affects the fitting of adjacent overlapping spectral peaks. To our knowledge, the extent of these possible interrelationships has never been assessed for in-vivo 31P MRS spectra.

A recognized limitation of in-vivo spectroscopy is the low sensitivity of 31P MRS and the low concentrations of 31P metabolites, which limit the spatial resolution of 31P MRSI. There is a potential for further improvement in both signal to noise and spectral resolution at higher magnetic field strengths. Second, because the MRSI technique utilizes phase encoding for spatial localization, there is some signal contamination from outside the selected VOX. To reduce contamination, a Hamming spatial filter was used, which, however, also decreased spatial resolution. Third, the use of different coils for proton MRI and 31P MRSI required re- moving the patient from the magnet to change coils, which lengthened the study and introduced the possibility of movement and misregistration of the MR and metabolite images. Construction of double-tuned high sensitivity coils in our laboratory should facilitate the combination of MR and 31P MRSI.

In summary, this study provides evidence for alterations in brain PDE and PCr in the frontal but not parietal lobes of patients with chronic schizophrenia using 3*P MRSI. These observations provide further support for altered frontal lobe functioning in schizophrenia.

This research was supported by a NARSAD Young Investigator Award (RFD); BRSG Grant S07-RRO5755 of the Division of Research Resources, NIH (RFD); UCSF Academic Senate Individual Investigator Grant (RFD); VA Research Associate Career Development Award (RFD): VA Research Career Scientist Award (GF); VA Medical Research Service (MWW); and the Philips Medical Systems (MWW).

DeLisi LE, Holcomb HH, Cohen RM, et a1 (1985): Positron emission tomography in schizophrenic patients with and without neuroleptic medication. J Cereb Blood Flow Metab

Farkas T, Wolf AP, Jaeger J, Brodie JD, Christman DR, Fowler JS (1984): Regional brain glucose metabolism in chronic schizo- phrenia. Arch Gen Psychiatry41 :293-300.

Fujimoto T, Nakano T, Takano T, et al (1992): Study of chronic schizophrenia using 31P chemical shift imaging. Pror SOC Magn Reson Med 1 : 1935.

Gattaz WF, Kollisch M, Thuren T, Virtanen JA, finnunen PKJ (1987): Increased plasma phospholipase A2 activity in schizo-

5 120 1-206.

510 BIOL PSYCHIATRY 1994;36:5O3-5 10

R.F. Deicken et d

phrenic patients: reduction after neuroleptic therapy. Biol Psy- chiatry 22:421426.

Gattaz WF, Hubner CK, Nevalainen TJ, Thuren T, Kinnunen PKJ (1990): Increased serum phospholipase A2 activity in schizo- phrenia: a replication study. Biol Psychiatry 28:495-501.

Geraud G , Arne-Bes MC, Guell A, Bes A. (1987): Reversibility of hemodynamic hypofrontality in schizophrenia. J Cereb Blood Flow Metab 7:9-12.

Glonek T, Kopp SJ, Kot E, Pettegrew JW, Harrison WH, Cohen MM (1982): P-31 NMR resonance analysis of brain: perchloric acidextract spectrum. JNeurochem 39:121&1219.

Horrobin DF, Manku MS, Hillman H, Iain A, Glen M (1992): Fatty acid levels in the brain of schizophrenic and normal controls. Biol Psychiatry 30:795-805.

Hugg JW, Matson GB, Twieg DB, Maudsley AA, Sappey-Marin- ier D, Weiner M W (1992): 3 1P spectroscopic imaging of nor- mal and pathological human brains. Magn Reson Imaging IO:227-243.

Kat0 T, Shioiri T, Takahashi S, Inubushi T (1991): Measurement of brain phosphoinositide metabolism in bipolar patients using in vivo31PMRS. JAffecriveDisord22:185-190.

Kat0 T, Takahashi S , Shioiri T, Inubushi T (1992): Relationship between brain phosphorous metabolism and brain lithium con- centration in manic-depressive illness detected by 3 1P and 7Li MRS. Proc SOC Magn Reson Med 1 : 1934.

Keshavan MS., Pettegrew JW, Panchalingam K, Kaplan D. Brar J, Campbell K (1 989): In vivo 3 1P nuclear magnetic resonance (NMR) spectroscopy of the frontal lobe metabolism in neuro- leptic naive first episode psychoses: preliminary studies. Schi- zophr Res 2: 122.

Kling A, Metter EJ. Riege WH, Kuhl DE (1986): Comparison of PET measurement of brain atrophy in chronic schizophrenia and depression. Am J Psychiatry 143: 175-1 80.

Maudsley AA, Twieg DB, Sappey Marinier D, et a1 (1990): Spin- echo 3 1P spectroscopic imaging in the human brain. Magn Reson Med 14:415-422.

Minshew NJ, Pettegrew JW, Payton JB, Panchalingam K (1989): Metabolic alterations in the dorsal prefrontal cortex of autistics with normal IQ. Ann Neuro126:438A2.

Oishi K, Zheng B, Kuo JF (1990): Inhibition of NA, K-ATPase and sodium pump by protein C kinase regulators sphingosine, lysophosphatidylcholine, and oleic acid. J Biol Chem 265:7C- 75.

Overall JE, Gorham DR (1962): The brief psychiatric rating scale. Psycho1 Rep 10:799-8 12.

Pangerl AM, Steudle A, Jaroni HW, Rufer R, Gattaz WF (1991):

Increased platelet membrane ly~~pho~phatidylcholine in schizophrenia. Biol Psychiatry 30837-840.

Petroff OAC, Richard JW, Behar KL, Alger JR, den Hollander JA, Shulman RG (1985): Cerebral intracellular pH by 31-p NMR spectroscopy. Neurology 35781-788.

Pettegrew JW, Keshavan M, Panchalingam K, et al(1991): Alter- ations in brain high-energy phosphate and membrane phospho- lipid metabolism in first-episode, drug-naive schizophrenics. Arch Gen Psychiatry 48:563-568.

Roberts JKM, Wade-Jardetzky N, Jardetzky 0 (1981): Intracellu- lar pH measurements by 3 1 P nuclear magnetic resonance: in- fluence of factors other than pH on 31P chemical shifts. Bio- chemistry 205389-5394.

Sheppard G, Manchanda R, Gruzelier J, et a1 (1983): "0-Positron emission tomography scanning in predominantly never treated acute schizophrenics. Lancet 2: 1448-1452.

Spitzer RL, Williams JBW (1985): The Structured Clinical Inter- view for DSM-Ill. New York: Biometrics Research Depart- ment, New York State Psychiatric Institute.

Stanley JA, Drost DJ, Williamson PC, Carf T, Rylett J, Merskey H (1 992): In vivo I'P MR spectroscopy study of schizophrenics at different stages of illness. Proc SOC Magn Reson Med 1 :756.

Szechtman H, Nahmias C, Gamett ES, et a1 (1988): Effect of neuroleptics on altered cerebral glucose metabolism in schizo- phrenia. Arch Gen Psychiatry 45523-532.

Van den Bosch H (1980): Intracdlular phospholipases. Biochem Biophys Acta 604: 19 1-246.

Volkow ND, Brcdie JD, Wolf AP, Angrist B, Russell J, Cancro R (1986): Brain metabolism in patients with schizophrenia be- fore and after acute neuroleptic administration. J Neurol Neur- osurg Psychiatry 49: 1 199-1202.

Volkow ND, Wolf AP, Van Gelder P, et a1 (1987): Phenomenolog- ical correlates of metabolic activity in 18 patients with chronic schizophrenia. Am J Psychiatry 144: 15 1-158.

Weinberger DR, Berman KF, Zec RF (1986): Physiological dys- function of dorsolateral prefrontal cortex in schizophrenia, I: regional cerebral blood flow (RCBF) evidence. Arch Gen Psy- chiurry 43: 114-124.

Williamson P, Drost D, Stanley J, Carr T, Morrison S, Merskey H ( 1991): Localized phosphorous 3 I nuclear magnetic resonance spectroscopy in chronic schizophrenic patients and normal controls. Arch Gen Psychiutty 48578. Letter.

Wolkin A, Jaeger J, Brodie J, et a1 (1985): Persistence of cerebral metabolic abnormalities in chronic schizophrenia as deter- mined by positron emission tomography. Am J Psychiatry 142:564-57 1.

- 24- stu

Per E Erlin -

Intrc Lithii recun theraI conce can b - From th

COI k r net1

Addresr Pha pen

Receive

8 1994